SIL/SDA participant plan reductions and cohort savings projections (Individual advocacy) | External submission

Submission 467

Submission to the

Senate Community Affairs Legislation Committee

Inquiry into the

National Disability Insurance Scheme Amendment

(Securing the NDIS for Future Generations) Bill 2026

Submitted by

Performl

Date: 28 May 2026

Contact: Loki Ball

Executive Summary

The National Disability Insurance Scheme Amendment (Securing the NDIS for Future Generations) Bill

2026 implements 10 reform measures projected by Treasury to reduce NDIS spending by $38.1 billion over the five years to 30 June 2030, and by approximately $185 billion over twelve years (Senate Order response MC26-010216, 26 May 2026, page 3).

This submission provides an analysis of three questions the Bill places before the Parliament. First, who stays in the NDIS and who leaves under the reform. Second, which measures the Treasury savings figure depends upon. Third, how the burden of the per-participant savings is distributed between cohorts — particularly between participants with very high support needs (those using Supported Independent Living or Specialist Disability Accommodation, hereafter “SIL/SDA”) and the rest of the scheme.

The submission has three findings.

Finding 1 — The reform removes 346,000 participants from the projected scheme by 30 June 2031. Treasury discloses that 241,000 of these are pre-1 January 2028 entrants who fail post-reform eligibility tests (Senate Order response, page 2). A further 105,000 participants are absent from the reformed projection without being explicitly named. This “inflow gap” — the difference between the no-reform population minus stated exits and the after-reform population — is likely the combined effect of reduced new-entrant flows and additional exits attributable to reform measures 1, 5, and 6. Government has not disclosed the composition of these groups.

Finding 2 — Treasury’s $38.1 billion savings figure is concentrated in four measures. Measure 2 (the reset of social, economic and community participation, and allied health and therapy budgets) accounts for $13.2 billion, measure 7 (the Objective Functional Capacity Test) for $9.3 billion, measure 5 (the New Framework) for $3.9 billion, and measure 4 (reassessment and unspent funds reforms) for $3.1 billion. These four measures deliver 78% of total savings. The remaining six measures together contribute $8.6 billion.

Finding 3 — Treasury has not disclosed how the per-participant savings burden falls between the SIL/SDA cohort and other participants. Two distributions are consistent with the published $38.1 billion envelope: one in which SIL/SDA participants face plan reductions of approximately 15% (with

Page 1

Submission 467

other participants at 13%), and one in which SIL/SDA participants are protected from measures 1, 2, and 5 and other participants face reductions of approximately 22%. The choice involves a transfer of approximately $3.4 billion per year between cohorts at the FY31 run rate. The trade-off and its implications are developed in Part 6.

These choices have material consequences for hundreds of thousands of participants, their families and carers, and the provider market that has developed to serve them. They are currently invisible to the Parliament, to participants, and to the providers being asked to plan for the post-reform environment. This submission’s recommendations focus on resolving that invisibility.

Recommendations

The Community Affairs Legislation Committee should report to the Senate on the following matters.

Recommendation 1 — Disclosure of modelling assumptions. Before the Bill proceeds to a vote, the Australian Government should publish the full cohort incidence assumptions underlying Treasury’s $38.1 billion savings projection. Specifically, the Government should disclose:

(a) the assumed split of each of the ten reform measures’ savings between SIL/SDA participants and other participants, expressed in dollars per annum for each year of the projection and as a percentage of each cohort’s no-reform spending;
(b) the demographic and functional composition assumed for the 241,000 stated exit cohort — by primary disability type, age band, current plan size band, and current functional assessment classification;
(c) the mechanism producing the 105,000-participant inflow gap — specifically, the assumed reduction in gross new-entrant flows attributable to each measure, additional exits from post-1 January 2028 entrants, and the assumed Thriving Kids cohort beyond the stated exits; and
(d) the per-participant plan budget changes implied for both cohorts in each year of the projection. Recommendation 2 — Clarity on SIL/SDA cohort treatment. The Australian Government should resolve the ambiguity currently surrounding the treatment of participants with very high support needs. Either the Bill or an associated instrument should firmly guarantee full protection of SIL/SDA participants from measures 1, 2, 3, and 5 — or the Government should transparently confirm the magnitude of per-participant impact those participants will face, publishing the cohort-specific impact in the same form as the headline savings projection and committing to monitor it during implementation. Participants, families, and providers cannot make safe planning decisions without knowing which path applies.

Recommendation 3 — Comprehensive public monitoring of reform impact. The Bill should require quarterly public reporting of reform impact, beginning at the date of commencement, covering:

(a) per-cohort plan budget changes for participants remaining in the scheme — average and distributional changes for the SIL/SDA cohort and for other participants, broken down by

Page 2

Submission 467

primary disability, age band, and category, permitting comparison of projected against actual impacts;

(b) exit volumes by primary disability, age band, current plan size band, and prior tenure, including what happens to exited participants (transition to Thriving Kids, other programs, or no formal support);
(c) provider market response indicators, including SIL capacity by region, SDA dwelling enrolment and pipeline, provider counts by service category, and disability support workforce metrics; and
(d) plan utilisation patterns by category, including distributional changes and patterns of unspent funds where these remain measurable post-reform. Reporting should be published in a form that permits independent scrutiny, including structured data

releases. The monitoring requirement should be embedded in the Bill rather than left to

administrative discretion.

These three recommendations are interdependent. Without disclosure, the Parliament cannot scrutinise the cohort incidence choice. Without resolving the cohort incidence choice, participants and providers cannot plan. Without monitoring, the Parliament cannot judge whether the implemented reform performs as projected.

Part 1 — The scheme as at 31 March 2026

1.1 Scale and composition The National Disability Insurance Scheme supported 774,456 active participants with approved plans as at 31 March 2026, with 18,530 new participants entering the scheme during the March 2026 quarter alone (NDIA Quarterly Report Q3 2025-26, Section 1.1). The scheme has grown from 534,655 participants in 2022 to 774,456 in 2026 — an increase of approximately 45% over four years.

Treasury projects total NDIS payments of $49.8 billion for the financial year ending 30 June 2026 (Senate Order response, page 3). Payments are concentrated in two categories. Core daily activities — which includes Supported Independent Living — accounted for $26.0 billion (49% of total payments) in the twelve months to 31 March 2026 (QR Q3 2025-26, Section 4.1). Social, economic and community participation accounted for approximately 24%, and capacity building allied health and therapy supports for 12%.

The average annualised plan budget across all active participants was $90,600 at 31 March 2026, up from $82,500 a year earlier — annual growth of 9.8% (QR Q3 2025-26, Figure 31). For participants continuing in the scheme across that period the average rose from $82,500 to $85,100, indicating that some of the headline growth reflects new entrants with higher initial plan sizes.

Participants are distributed unevenly across age and disability. Autism is the largest single primary disability category. Children aged 0-14 account for approximately 39% of all participants, with participation rates peaking at 11.5% of the Australian population at age 6 (QR Q3 2025-26, Section

Page 3

Submission 467

1.2). Psychosocial disability and intellectual disability are the next largest disability categories after autism.

1.2 The SIL/SDA cohort Within the broader participant population, a distinct group of participants with very high support needs uses Supported Independent Living arrangements, Specialist Disability Accommodation, or both. These are participants whose disability and functional requirements are sufficient that they require support with daily life, purpose-built accessible housing, or both.

At 31 March 2026, the SIL sub-cohort comprised 36,808 participants, with average plan budgets of $488,600 — approximately 5.4 times the scheme-wide average plan budget of $90,600 (QR Q3 2025 26, Table 10 and Figure 31). The SDA sub-cohort comprised 16,263 participants using SDA dwellings plus a further 9,370 participants eligible for SDA but not yet using it (QR Q3 2025-26, Table 11). Combining these sub-populations and adjusting for overlap (many SDA users also receive SIL supports), the overall SIL/SDA cohort comprises approximately 41,300 participants — 5.3% of the scheme.

The cohort’s supports — overnight staffing, accessible accommodation, complex care— cannot be substituted within the scheme by reductions in other budget components. A reform measure that reduces a SIL/SDA participant’s plan by 15% is, in concrete terms, a reduction in supported living hours or specialised accommodation funding due to the flexible use of core and capacity building funding for essential care. These structural features make the cohort incidence of reform measures a substantive design question.

Part 2 — The status quo trajectory to 30 June 2031

2.1 The Treasury participant projection Treasury’s modelling, disclosed in response to Senate Orders 503-508, projects that under the status quo — i.e. in the absence of the reform measures — the NDIS would grow from 739,414 participants at 30 June 2025 to 944,000 participants at 30 June 2031 (Senate Order response, page 2). This is net growth of approximately 205,000 participants over six years, or roughly 34,000 net new participants per year.

The growth profile is steady. Treasury projects 783,000 participants at 30 June 2026, 817,000 at 30 June 2027, 842,000 at 30 June 2028, 876,000 at 30 June 2029, 910,000 at 30 June 2030, and 944,000 at 30 June 2031. The implied annual net growth rate declines from 5.9% in FY25-26 to 3.7% by FY30 31, consistent with the scheme approaching its long-term steady-state participation rate.

These figures are Treasury’s stated projection, and form the baseline against which the reform measures are evaluated. They reconcile to NDIA Q3 2025-26 actuals at the starting point (774,456 participants at 31 March 2026 with a trajectory toward 783,000 by 30 June 2026).

2.2 The implied payments trajectory

Page 4

Submission 467

Treasury’s no-reform payments trajectory is set out in two forms across the Senate Order response. The first, from the 2025-26 Mid-Year Economic and Fiscal Outlook (MYEFO), projects payments rising from $49.1 billion in FY25-26 to $66.4 billion in FY29-30 — five-year total of $287.2 billion. The second, derived by adding the measure savings from page 4 to the post-reform totals on page 3, gives $49.8 billion in FY25-26 rising to $70.9 billion in FY29-30.

Extrapolating the second series, the no-reform projection reaches approximately $76 billion in FY30

Over the seven years from FY25-26 to FY31-32, the scheme would deliver approximately $390 billion of payments under the no-reform scenario.

The growth rate of payments under no-reform substantially exceeds the growth rate of participants. Participants grow at approximately 5% per annum on average; payments grow at approximately 8 10% per annum. The differential reflects the average plan budget rising over time, which itself reflects two effects: (a) the per-participant cost of supports is rising due to wages, indexation, and complexity; and (b) the participant mix is shifting toward higher-needs participants as the scheme matures, as lower-needs participants tend to exit and longer-tenured participants tend to acquire larger plans.

2.3 The SIL/SDA cohort trajectory under status quo Annualised SIL payments totalled $16.4 billion in the twelve months to March 2026, up from $13.4 billion two years earlier. Total SDA payments reached $539 million in the year to March 2026, up from $290 million two years earlier — annual growth of 36%. Combined SIL/SDA spending is approximately $19.0 billion, or approximately 37% of total scheme payments.

Under the no-reform projection, the SIL/SDA cohort grows from approximately 41,300 participants at March 2026 to 51,654 at 30 June 2031 — net growth of approximately 10,400 participants, or roughly 2,000 per year. The cohort’s share of total scheme participants remains approximately 5.5% throughout the projection.

Cohort spending, however, grows substantially faster than headcount. Annualised SIL/SDA payments rise from approximately $19.0 billion at March 2026 to approximately $30.1 billion at 30 June 2031 — annual growth of approximately 9.7%. The average plan size for SIL/SDA participants rises from approximately $460,000 to approximately $583,000 over the same period, an increase of 27% in nominal terms.

The cohort’s share of total scheme payments grows from approximately 37% at March 2026 to approximately 39% at 30 June 2031. This is not because SIL/SDA participants grow as a share of headcount, but because their average plan grows faster than the scheme-wide average. Three dynamics drive this growth. First, more participants are using their SDA funding as the new build SDA market grows — enrolment was up 25% annually over the two years to March 2026 (QR Q3 2025-26, Section 4.5). Second, the cost of SIL supports has been rising at approximately 6% per annum, reflecting wage increases for disability support workers, indexation under the SCHADS award, and growth in overnight and complex care components. Third, participants who have been receiving SIL or SDA supports for several years tend to see their plans grow over time, as support needs are reassessed upwards, additional services are added, and care complexity increases with age.

Page 5

Submission 467

2.4 The fiscal context The status quo trajectory is the source of the fiscal pressure that the reform and Bill responds to. Cumulative scheme payments over the seven years from FY25-26 to FY31-32 would be approximately $390 billion under no-reform, against Treasury’s twelve-year no-reform projection that exceeds $900 billion (Senate Order response, page 3, MYEFO row 12-year total of $913.5 billion).

Cohort-level fiscal pressure is also concentrated. By 30 June 2031, the no-reform trajectory implies the SIL/SDA cohort would account for approximately $30 billion of annual scheme spending — roughly 39% of total payments going to 5.5% of participants. The remaining 892,346 participants would account for approximately $46 billion in annual spending, with an average plan of approximately $52,000 per participant.

These structural features inform the design of the reform measures the Bill implements. The next Part describes what those measures are.

Part 3 — The reform package

3.1 The ten measures The reform package implements ten distinct measures, projected to deliver $38.1 billion in savings on a fiscal balance basis over the five years from FY25-26 to FY29-30 (Senate Order response, page 4). On an underlying cash basis the equivalent figure is $37.8 billion.

The measures and their five-year savings envelopes are as follows.

Measure 1 — Strengthen guidance around what is reasonable and necessary: $2.9 billion.
Measure 2 — Reset of core social, economic, and community participation and capacity building allied health and therapy budgets: $13.2 billion.
Measure 3 — Plan management and support coordination reforms: $1.4 billion.
Measure 4 — Tightening criteria around unscheduled reassessment requests, ending plan rollovers, and stopping unspent funds: $3.1 billion.
Measure 5 — Lower growth due to implementation of the New Framework: $3.9 billion.
Measure 6 — Access changes to enable Thriving Kids rollout: $0.5 billion.
Measure 7 — Introduce an Objective Test of Substantially Reduced Functional Capacity: $9.3 billion.
Measure 8 — Differentiated pricing for unregistered providers: $2.6 billion.
Measure 9 — Mandatory registration for high-risk providers: $0.2 billion.
Measure 10 — Making the Minister the decision-maker on pricing, and related fraud measures: $0.9 billion. The savings ramp is steeply backloaded. The measures collectively deliver no fiscal balance impact in FY25-26, $2.0 billion in FY26-27, $7.5 billion in FY27-28, $11.9 billion in FY28-29, and $16.7 billion in

Page 6

Submission 467

FY29-30. Approximately three-quarters of the cumulative savings accrue in the final two years of the modelled window, reflecting the time required for reassessment cycles, eligibility tests, and budget resets to propagate through the participant population. The four measures that dominate the savings profile — measures 2, 7, 5, and 4 — are examined in Part 5.

3.2 Two structural mechanisms The ten measures operate through two structurally different mechanisms.

The first mechanism is participant exit. Measures 6 (Thriving Kids) and 7 (Objective Functional Capacity Test) operate by removing participants. Together these measures account for $9.8 billion of savings over five years, and Treasury discloses that 241,000 participants who entered the scheme before 1 January 2028 will be exited under this mechanism by 30 June 2031 (Senate Order response, page 2).

The second mechanism is per-participant budget and payment reduction. Measures 1, 2, 3, 4, 5, 8, 9, and 10 — totalling $28.3 billion in savings — operate by reducing the size of plans for participants who remain in the scheme. These measures take various forms: budget resets at reassessment (measure 2); growth rate caps (measure 5); tightened utilisation rules (measure 4); price reductions for specific provider types (measures 8 and 9); and tighter reasonable-and-necessary guidance (measure 1).

The distinction between these two mechanisms matters because they fall differently across the participant population. Exit measures are concentrated by their design in participants near the eligibility margin — predominantly children, participants with milder functional support needs, and participants with disabilities whose support needs the new tests would not validate at current plan sizes. By contrast, per-participant budget measures fall across the remaining scheme population, including across high-needs cohorts who would not be exited but whose plans would still be reduced.

Understanding this structural distinction and its varied impact across different cohorts is important.

3.3 The reformed scheme at 30 June 2031 Applying the reform measures to the no-reform trajectory produces an after-reform NDIS of materially different scale.

Participants reduce from 944,000 (no-reform) to 598,000 (after-reform) at 30 June 2031 — a reduction of 346,000 participants, or 36.7% (Senate Order response, page 2). Of this reduction, 241,000 are explicitly disclosed by Treasury as exits via measure 7’s eligibility test. The remaining 105,000 are absent from the reformed projection without being explicitly named — this submission terms this the “inflow gap” and addresses its likely composition in Part 4.

Scheme payments reduce correspondingly. Against the no-reform trajectory of approximately $76 billion in FY30-31, the post-reform trajectory is approximately $54 billion in FY29-30 (the last year for which Treasury publishes the figure). Extrapolating one year forward, the reformed scheme spends

approximately  $56-58  billion  in  FY30-31. The  difference between  the two  trajectories —

approximately $20 billion per year at run-rate — is the annual fiscal saving the reform delivers, with the cumulative savings reaching $38.1 billion over the five years from FY25-26 to FY29-30.

Page 7

Submission 467

The reformed scheme is therefore approximately 64% the size of the no-reform scheme by headcount, and approximately 74% by spending. Per-participant average payments under the reform rise from approximately $80,000 (no-reform, $76bn / 944k) to approximately $94,000 (reform, $56bn / 598k) — a 17% increase in nominal terms. This rise reflects the participants who exit have lower-than average plan sizes, raising the average for those who remain.

Part 4 — Who stays, who goes

4.1 The participant gap By 30 June 2031, the reform produces a scheme of 598,000 participants — 346,000 fewer than the no-reform projection of 944,000 (Senate Order response, page 2). This gap builds progressively across the projection: zero through 30 June 2027, 53,000 by 30 June 2028, 166,000 by 30 June 2029, 278,000 by 30 June 2030, and 346,000 by 30 June 2031.

Treasury’s disclosure separates this gap into one stated component and one unstated component. The stated component, identified on page 2 of the Senate Order response as “Entered scheme before 1 Jan 28 and assumed to be exited”, totals 33,000 participants at 30 June 2028, 125,000 at 30 June 2029, 205,000 at 30 June 2030, and 241,000 at 30 June 2031. The unstated component — the difference between the no-reform projection minus the stated exits and the after-reform total — totals 20,000 at 30 June 2028, 41,000 at 30 June 2029, 73,000 at 30 June 2030, and 105,000 at 30 June 2031. This submission terms the unstated component the “inflow gap.” Treasury has not named or explained it in the published material.

4.2 The 241,000 stated exit cohort The stated exits are pre-1 January 2028 entrants who, under the reform and Bill, will be subject to reassessment against the new Objective Test of Substantially Reduced Functional Capacity (measure

and the access changes that enable the Thriving Kids rollout (measure 6). Treasury attributes $9.8 billion of cumulative savings to these two measures over the five years from FY25-26 to FY29-30.

The Treasury document does not specify which participants are projected to exit, but the demographic and functional composition can be inferred from two anchors. First, the cumulative average annualised cost per exited participant implied by Treasury’s annual savings ramp falls from approximately $61,000 in FY27-28 to approximately $31,000 by FY30-31. Second, the savings can only be delivered if the exits are drawn from non-SIL/SDA participants, as very high needs SIL/SDA participants would by definition pass any Substantially Reduced Functional Capacity test and would not exit under Thriving Kids.

Given these anchors, the exits are concentrated in participant profiles with lower-than-average plan budgets and milder functional needs. Building a cell-level model that reconciles to Treasury’s cumulative cost-per-exit anchor and the documented exit volumes produces an exit profile dominated by autism (approximately 150,000 of the 241,000 exits, 62%), developmental delay and global

developmental  delay  (approximately  57,000), and  intellectual  disability with  milder  profiles

(approximately 9,000). The exits are heavily concentrated by age: approximately 96,000 from the 0-8

Page 8

Submission 467

age band, approximately 67,000 from the 9-14 age band, and the remainder spread across the 15+ age bands.

This profile is consistent with the design of measures 6 and 7, but the actual selection of which participants exit will be determined by the operational application of the Objective Functional Capacity Test. Treasury has not disclosed how the test will be applied at the individual level, nor what the projected pass-rate is for each demographic and functional profile within the eligible population.

4.3 The inflow gap The 105,000-participant reduction not explained by stated exits — the inflow gap — is not addressed in the Treasury estimates. Three plausible mechanisms can produce it. The first is reduced new entrant flows under the reformed access regime: tighter reasonable and necessary guidance (measure

and the New Framework (measure 5) would make some prospective applicants ineligible or result in smaller initial plans, reducing the rate at which new participants join the scheme. The second is additional exits from post-1 January 2028 entrants — participants who join the scheme after the reform commences but who fail the new tests at their first reassessment. The third is Thriving Kids exits beyond those captured in the stated cohort.

The most likely explanation is a combination of all three, with reduced new-entrant flows representing the largest component. Under the no-reform projection, gross new entrants run at approximately 80,000 per year against natural exits of approximately 45,000 per year, producing net growth of approximately 35,000. The reform’s tightened access pathways could plausibly reduce gross new entrant flows by approximately 30,000 per year by FY29-30 — a 38% reduction — which would account for the bulk of the inflow gap.

The composition of the inflow gap matters substantively. If the gap is primarily reduced new-entrant flows, it represents prospective applicants who do not gain access to the NDIS — participants whose disabilities and support needs are real, but whose access pathway has been narrowed. If it is primarily additional exits, it represents existing participants whose plans are not renewed at reassessment. The policy and human consequences differ between these two cases.

4.4 The exit profile in summary Combining the stated exits and the inflow gap into a single view of who is removed from the projected NDIS, the composition is as follows.

By disability, the largest cohorts affected are autism (approximately 216,000 fewer participants by 30 June 2031, or 53% of the no-reform autism population), developmental delay (approximately 67,000

fewer, 82%),  global developmental delay (approximately 17,000, 72%),  intellectual  disability

(approximately 12,000, 12%), and psychosocial disability (approximately 9,000, 13%). Cohorts with high and consistent functional needs — cerebral palsy, down syndrome, spinal cord injury, multiple sclerosis — show much lower reductions of 4-6%, reflecting that their members would not exit under the new functional tests and have lower exposure to inflow tightening.

Page 9

Submission 467

By age, the reduction is heavily child concentrated. The 0-8 age band reduces by approximately 145,000 participants (71% of the no-reform population in that band). The 9-14 age band reduces by approximately 98,000 (52%). Together, children aged 0-14 account for 243,000 of the 346,000 total reductions — 70% of the participant gap. Adults aged 35 and over show reductions of 11-16%, largely concentrated in the inflow gap rather than exits.

These profiles emerge from the policy structure of the reform measures rather than from any explicit cohort-selection decision. The Objective Functional Capacity Test, by its design, will catch participants whose functional profile sits near the eligibility threshold, and these are predominantly children with developmental or behavioural diagnoses and adults with milder presentations. The Thriving Kids access changes are explicitly designed around the 0-8 age band. The inflow gap concentrates in age bands where new entrants are most numerous — children — and in disabilities where eligibility is being tightened.

4.5 What exiting the NDIS means For the 346,000 participants who would have been in the no-reform projection but are not in the after-reform projection, the consequences are not addressed in detail in the Treasury estimates. Three observations are relevant.

First, for the 241,000 stated exits, the consequence is the loss of all NDIS-funded supports. Plans currently averaging $30,000 to $60,000 per year — funding personal care, behavioural support, therapy, community participation, transport, and assistive technology — would cease at the point of reassessment failure. The participants concerned would no longer be NDIS participants.

Second, for the children projected to be moved into the Thriving Kids program rather than the NDIS, the supports available will depend on the design and funding of that program. The detail of how Thriving Kids will operate, what supports will be available, and on what funding scale, is still largely unknown.

Third, for the inflow gap of 105,000 participants who would have entered the NDIS under the no reform projection but who do not under the reform, the consequence is non-entry to the scheme. Their disabilities and support needs do not disappear; they simply do not gain access to scheme funded supports.

These three groups together represent a structural reorganisation of who is and who is not eligible for

funded  disability support  in Australia, of a scale not previously contemplated. Whether the

reorganisation is the right one is a question of policy. The question this submission raises is whether the Parliament should be deciding the matter without the disclosure of who, demographically and functionally, is being removed from the projected scheme.

Page 10

Submission 467

Part 5 — The big levers driving the savings

5.1 Four measures, 78% of the savings Treasury’s $38.1 billion savings figure is not evenly distributed across the ten reform measures. Four measures account for $29.5 billion — 78% of the total. The remaining six measures collectively contribute $8.6 billion, with the smallest (measure 9, mandatory registration of high-risk providers) contributing $0.2 billion.

The four major measures are as follows.

Measure 2 — Reset of core social, economic and community participation and capacity building allied health and therapy budgets ($13.2 billion). This is the single largest source of savings, accounting for 35% of the total. The measure operates by reducing the size of these specific budget components within participant plans. Treasury has not published the rate of reduction applied or the scope of plans affected — whether the reduction applies to all plans immediately, only to plans on reassessment, or to a subset of high-utilisation plans within these categories. Across the NDIS, these two support categories represent approximately $18 billion of annualised spending at the FY30-31 no reform run-rate, so the measure’s savings represent a reduction of approximately 24% of the total spend in these categories by FY29-30, building to approximately 25% at the FY30-31 run-rate.

Measure 7 — Objective Test of Substantially Reduced Functional Capacity ($9.3 billion). This is the eligibility test that produces the 241,000 stated exits. The measure delivers savings by removing participants from the NDIS rather than by reducing per-participant budgets. Treasury has not disclosed the parameters of the test — what functional thresholds it applies, what evidence is required, or what the projected pass-rate is for different disability and age profiles.

Measure 5 — Lower growth due to implementation of the New Framework ($3.9 billion). This measure delivers savings by constraining the growth rate of participant plan budgets, rather than by cutting plans directly. Treasury has not disclosed the growth rate assumed under the New Framework

or how  it differs from the no-reform growth path. The measure  is particularly back-loaded —

delivering 64% of its savings ($2.5 billion of $3.9 billion) in FY29-30 alone — consistent with cumulative effects of constrained growth compounding over time.

Measure 4 — Tightening reassessment criteria and ending plan rollovers and unspent funds ($3.1 billion). This measure operates through three sub-mechanisms: limiting unscheduled reassessment requests, ending the practice of unspent funds rolling over between plans, and removing carry forward provisions. Treasury has not disclosed the relative contribution of each sub-mechanism or whether they apply uniformly across the participant population.

5.2 What Treasury has and has not disclosed For each of the four major measures, the savings envelope is published, the annual ramp profile is published, and the underlying mechanism is described in summary form. What is not published is the operational specification — which participants are affected, by how much, and when. Specifically, Treasury has not disclosed:

Page 11

Submission 467

whether SIL/SDA participants are subject to measure 2’s budget reset, and whether the reduction applies uniformly across geographies (e.g. remote vs urban) or disability types (e.g. autism vs intellectual disability);
which 241,000 of the projected 838,000 pre-1 January 2028 entrants exit under measure 7, and on what disability or functional profile (though the cumulative cost-per-exit anchor implies exits are skewed to lower-cost participant profiles);
whether measure 5’s growth-rate constraint applies uniformly to all participants and all categories, or has cohort-specific or category-specific application; and
how measure 4’s reassessment-tightening will be applied, whether SIL/SDA participants are subject to the same reassessment criteria as other participants, and how the rollover and unspent-funds provisions interact with the cohort that has historically had near-full plan utilisation.

5.3 How the measures interact The measures do not operate independently. Measure 7’s exits remove the participants whose plans would otherwise have been affected by measure 2’s budget reset, measure 4’s reassessment tightening, and measure 5’s growth constraint. The order in which the measures take effect therefore determines what the remaining participant population looks like.

Treasury’s modelling implicitly assumes a sequencing in which exits propagate ahead of, or in parallel with, per-participant budget reductions. If the exits are concentrated in lower-cost participant profiles (as the data anchors imply), then the participants remaining in the NDIS after the exit phase are the higher-cost participants whose plans the budget measures then act upon. The effect is that the per participant budget measures fall more heavily on remaining higher-cost cohorts than they would have done on the no-reform participant mix.

For SIL/SDA participants, the exit measures do not apply. Measure 7 (the Objective Functional Capacity Test) is designed to identify participants whose functional capacity has not been substantially reduced — a test that SIL/SDA participants meet, given their very high support needs and substantially reduced functional capacity. Measure 6 (Thriving Kids) targets the 0-8 age band, which contains effectively no SIL/SDA participants. The post-reform plan size of SIL/SDA participants is therefore determined by the operational design of the eight per-participant measures (measures 1, 2, 3, 4, 5, 8, 9, and 10).

Part 6 — The choice: protect or distribute

6.1 The structure of the choice By 30 June 2031, the reform measures are projected to deliver $20.3 billion in annual savings at the run-rate, of which $7.0 billion arises from the participant exits already examined in Parts 3 and 4, and the remaining $13.3 billion arises from per-participant budget and payment reductions across measures 1, 2, 3, 4, 5, 8, 9, and 10. The $13.3 billion of per-participant savings must fall on the 598,000 participants who remain in the scheme.

Page 12

Submission 467

Of those 598,000 participants, approximately 51,654 are projected to be in the SIL/SDA cohort, and 546,346 are projected to be outside it. The cohort-incidence question is how the $13.3 billion of per participant savings is allocated between these two groups.

Treasury’s modelling envelope is consistent with two materially different distributions of the burden. Both scenarios deliver the same total $38.1 billion of cumulative savings over the five years to FY29 30, and both reconcile to the participant projections on page 2 of the Senate Order response. They differ only in which cohort bears which share of the per-participant burden.

The Treasury document does not disclose which distribution is assumed.

6.2 Scenario A — Cohort-neutral distribution The first scenario assumes each measure is applied to both cohorts in proportion to their share of no reform spending in the categories the measure targets. This assumes uniform application across the participant population — for example, a budget reset that applies to all plans.

The SIL/SDA cohort, with no-reform payments of approximately $30.1 billion at FY30-31, absorbs approximately $4.7 billion in savings. This corresponds to an average plan budget reduction of approximately 15.2% per participant — from approximately $601,500 to approximately $510,000 per year. The reduction is concentrated in measure 2 ($1.4 billion, the budget reset), measure 5 ($1.6 billion, growth rate caps under the New Framework), measure 1 ($433 million, reasonable-and necessary guidance), and measure 4 ($700 million, reassessment tightening).

The non-SIL/SDA cohort that remains in the NDIS, with no-reform payments of approximately $46.4 billion at FY30-31 absorbing 892,346 participants but reducing to 546,346 after the exits, absorbs approximately $15.6 billion of savings — comprising the $7.0 billion from exits and $8.6 billion from per-participant measures. The exit savings reduce headcount; the per-participant savings reduce plan sizes for those remaining. The average plan budget for participants remaining in the cohort falls from approximately $70,300 (no-reform) to approximately $60,900 — a reduction of approximately 13.4%.

The headline outcome of Scenario A is therefore that both cohorts face approximately similar percentage reductions in their plan budgets — 15.2% for SIL/SDA, 13.4% for kept non-SIL/SDA. The reform burden is broadly proportional across the participant population.

6.3 Scenario B — SIL/SDA cohort protection The second scenario assumes that SIL/SDA participants are explicitly protected from measures 1, 2, and 5 — the three large per-participant measures. SIL/SDA participants remain subject to measures 3, 4, 8, 9, and 10.

Under this scenario, the SIL/SDA cohort absorbs approximately $1.3 billion in savings at the FY30-31 run-rate, against the $4.7 billion under Scenario A. This represents an average plan budget reduction of approximately 4.2% per participant — from approximately $601,500 to approximately $576,200 per year.

The savings shifted away from SIL/SDA — approximately $3.4 billion per year at the FY30-31 run-rate — must be recovered from the non-SIL/SDA cohort to keep the Treasury envelope intact. The non

Page 13

Submission 467

SIL/SDA cohort therefore absorbs approximately $19.0 billion in savings under Scenario B against $15.6 billion under Scenario A. The average plan budget for kept non-SIL/SDA participants falls from approximately $70,300 to approximately $54,600 — a reduction of approximately 22.3%.

The headline outcome of Scenario B is therefore that the burden is sharply concentrated. SIL/SDA participants face a 4.2% reduction; non-SIL/SDA participants who remain in the scheme face a 22.3% reduction.

6.4 The trade-off in dollar terms The dollar value of the cohort choice is straightforward.

For SIL/SDA participants, the move from Scenario A to Scenario B is worth approximately $66,200 per participant per year on average — the difference between an average plan of $510,000 and an average plan of $576,200. For a cohort of 51,654 participants, this aggregates to approximately $3.4 billion per year of additional supports that would be preserved under protection of the very high needs cohort of SIL/SDA participants.

For non-SIL/SDA participants who remain in the NDIS, the move from Scenario A to Scenario B costs approximately $6,300 per participant per year on average — the difference between an average plan of $60,900 and an average plan of $54,600. For a cohort of 546,346 participants, this aggregates to approximately $3.4 billion per year of additional supports forgone under cohort protection.

The qualitative character of the trade-off is asymmetric. For SIL/SDA participants, the $66,200 per participant impact corresponds to changes in supported living hours, accommodation funding, or critical disability supports that cannot be easily substituted. For non-SIL/SDA participants, the $6,300 per-participant impact corresponds to plan reductions that vary substantially across diagnoses and ages — falling more heavily on adults with psychosocial or intellectual disability whose plans tend to be larger, and more lightly on younger participants with smaller plans.

6.5 Why this choice is currently invisible

The cohort incidence choice  is consequential.  It determines whether approximately 51,654

participants face plan reductions of 4% or 15%, and whether approximately 546,346 participants face plan reductions of 13% or 22%. It involves the distribution of approximately $6.4 billion of reform burden between cohorts over the five-year Treasury window, and approximately $17 billion if extrapolated across the full twelve-year savings horizon.

Treasury has not disclosed which scenario it has modelled. The published material describes the measures, sets out the annual savings profile, and gives the participant trajectories. It does not disclose the per-measure cohort split assumed, the per-participant plan reduction implied for either cohort, or the operational basis on which measures will be applied to SIL/SDA participants.

A reform that explicitly redistributes approximately $6 billion of fiscal pressure between identifiable participant cohorts is properly a matter for debate. The operational consequences for participants, families, and providers are developed in Part 7.

Page 14

Submission 467

Part 7 — The cost of uncertainty The cohort incidence question is not only a question of fiscal distribution. It is also a question of operational consequence. The participants, families, carers, and providers whose decisions will be affected by the reform face a sustained period of uncertainty about what the reform will mean in practice.

7.1 Participants, families, and carers For the approximately 51,654 SIL/SDA participants and their families, the question is whether the supports their living arrangements depend on will continue at current levels or be reduced. For an individual SIL participant whose plan is at the cohort average of approximately $488,600 at March 2026, the difference between the two scenarios in Part 6 is approximately $54,000 per year in supported living and care funding.

For  the  approximately 546,346  non-SIL/SDA  participants who remain  in  the scheme,  the

corresponding uncertainty is whether their plan reduction will be moderate or sharp. For a participant on a $40,000 plan, the difference between the two scenarios is approximately $3,600 per year; for a participant on a $90,000 plan, the difference is approximately $8,100 per year.

For families and informal carers, these differences translate into questions about how much unpaid care will be required to bridge gaps between funded supports and continuing need. The uncertainty about which scenario applies, and the absence of a clear timeline for resolving it, encourages a pattern of pre-emptive risk management — reducing reliance on NDIS-funded supports, increasing informal care commitments, and reducing labour force participation by carers.

For the 346,000 participants in the no-reform projection who are not in the after-reform projection, the uncertainty is more fundamental. They do not yet know whether they will be exited under measure 7, captured in the inflow gap, or remain in the NDIS at all.

7.2 The provider market The NDIS provider market has been built around assumptions of growth and need. SIL providers have built operational scale on the basis of expected continued growth in the SIL participant base from 36,808 at March 2026 to approximately 47,000 by 2031. SDA developers have entered an investment pipeline that delivered a 25% per annum increase in enrolled dwellings over the past two years on the same expectation.

The reform may change that trajectory materially. If Scenario A is operative, payments to providers among SIL participants fall by approximately 15% by FY30-31, against an expectation of continued growth. Under Scenario B, the reduction is approximately 4%. The two scenarios imply substantially different capacity decisions and workforce levels for the provider market.

Without disclosure of which scenario is assumed, providers are required to make decisions under uncertainty about their post-reform participant profile. The rational response to such uncertainty in a thin-margin sector is to defer investment, reduce workforce expansion, and in some cases withdraw capacity. These responses, taken collectively, risk producing a contraction of supply ahead of the

Page 15

Submission 467

operational design of the reform — a contraction that would affect participants regardless of which scenario is ultimately operative. The Department of Health, Disability and Ageing and the National Disability Insurance Agency are necessarily working to an assumed cohort incidence in their operational planning; providers serving SIL/SDA participants are necessarily making investment, capacity, and workforce decisions on the basis of some assumption about what their participants’ plans will look like after 1 April 2027 when the New Framework commences.

7.3 The compounding cost of opacity The cost of uncertainty rises with time. Each quarter in which the operational design of the reform remains undisclosed adds to the cumulative volume of pre-emptive decisions made by participants, families, carers, and providers.

The  Bill  is scheduled for parliamentary consideration in mid-2026, with the New Framework

commencing on 1 April 2027 and major reform measures taking effect through 2030. The window between the Bill becoming law and the reform measures taking full effect is approximately three years. Across that window, the volume of decisions that depend on knowing how the reform applies to particular cohorts and categories will accumulate.

Three forms of decision are particularly affected.

Investment  decisions  in  capacity-building  infrastructure — SDA  dwellings,  supported  living

arrangements, specialised workforce training — have lead times of two to five years. Decisions made now are decisions that will deliver capacity around the time the reform takes effect. They depend on knowing what the post-reform demand profile looks like.

Workforce decisions — both individual decisions about whether to remain in the sector and organisational decisions about hiring and training — operate on shorter timeframes but are highly sensitive to expected demand. Workforce attrition that occurs in the period of uncertainty is unlikely to be recovered when clarity arrives.

Participant and family planning decisions — including decisions about housing transitions, returning to or remaining in the workforce, and the structuring of informal care arrangements — are deeply individual but are similarly difficult to reverse.

Each of these decision types is being made now, on the basis of incomplete information. The cost of uncertainty is the cumulative value of decisions that would have been made differently with better information. That cost rises with each month that the operational design of the reform remains undisclosed.

This is the substantive case for prompt and comprehensive disclosure of the modelling assumptions, resolution of the cohort incidence choice, and commitment to public reform monitoring, as set out in the Recommendations at the start of this submission.

Page 16

Submission 467

Methodology Appendix

This appendix documents the data sources, projection methods, cohort incidence modelling, exit selection model, and limitations underlying the analysis in this submission.

A.1 Data sources The analysis draws on three primary sources.

Senate Order response MC26-010216, dated 26 May 2026. A four-page response from the Hon Jenny McAllister, Minister for the NDIS, to the President of the Senate, providing the consolidated NDIS modelling response to Senate Orders 503-508 moved by Senator Steele-John and agreed on 14 May

The response contains: (a) participant projections to 30 June 2031 under pre-reform and post- reform scenarios; (b) annual NDIS payments by FY across 2025-26 MYEFO and 2026-27 Budget contexts; and (c) per-measure annual savings on a fiscal balance basis for each of the ten reform measures across FY25-26 to FY29-30.

NDIA Quarterly Report Q3 2025-26 (March 2026 quarter), released May 2026. The 112-page Quarterly Report and its supplements provide actual data as at 31 March 2026 on participant numbers, demographic composition, plan budgets and payments by support category, SIL and SDA cohort detail, provider market metrics, and outcome indicators. The submission draws primarily on Sections 1 (participants), 4 (providers and the growing market), and Tables 8, 10, and 11.

Specialist Disability Accommodation Pricing Review 2022-23 Final Report (National Disability

Insurance Agency, June 2023). The review sets out the published SDA payments trajectory to approximately $1.8 billion by 2033, with the SDA participant count growing to over 30,000 over the same period. The trajectory is decomposed into three drivers: greater use of new-build SDA stock (with higher prices per dwelling than legacy stock), reduction in use of existing legacy SDA dwellings as they are phased out, and growth in both participant numbers and aggregate payments. Actual SDA payments were $539 million in the year to March 2026. The submission draws on these data to anchor SDA component growth within the SIL/SDA cohort projection.

Performl’s integrated care data asset. A longitudinal dataset by small-area and cohort of NDIS payment and plan data extending back to 2019, structured at the line-item level and reconciled to NDIA-published aggregates. The asset combines payment data by support category and line item with participant cohort attributes (primary disability, age band, geography, SIL/SDA status) and provider side metrics. It is used in this submission to establish the no-reform projection’s category-level mix, the per-cohort spending shares within each budget category that underpin the cohort incidence modelling in Scenario A, the disability-by-age cell distribution used in the exit selection model, and the SDA and SIL trajectories that anchor the cohort-level growth path. The dataset’s seven-year span permits identification of structural growth trends distinct from period-specific effects and supports the back-cast reconciliation of quarterly values against NDIA quarterly reporting.

Page 17

Submission 467

A.2 Projection architecture The submission’s no-reform projection is built as a multi-layer cube spanning 22 quarters from March 2026 to June 2031. The cube has five layers:

Layer 1 — National totals, with sub-group breakdown by SIL/SDA cohort and the rest of the NDIS.
Layer 2 — Age × Disability dimension, comprising 10 age bands and 18 primary disability categories (180 cells).
Layer 3 — Support Class (Core, Capacity Building, Capital).
Layer 4 — Support Category (16 categories, split between SIL/SDA and non-SIL/SDA cohort within each).
Layer 5 — Cell-level detail at the intersection of Category × Class × Disability × Age × Cohort, providing the operational basis for cohort-specific modelling. Each layer reconciles to Layer 1 by construction. The cube is anchored to the Treasury participant projection at each June-end FY anchor and to NDIA quarterly actuals at the March 2026 starting point. Interior quarters are populated by linear interpolation between anchors with growth-rate scaling derived from cube structure.

A.3 Calibration and reconciliation The cube reconciles to the Treasury modelling at each June-end anchor on both participants and payments.

Participant reconciliation. The no-reform participant trajectory matches the Treasury page 2 projection exactly: 783,000 at Jun 2026, 817,000 at Jun 2027, 842,000 at Jun 2028, 876,000 at Jun 2029, 910,000 at Jun 2030, and 944,000 at Jun 2031.

Payments reconciliation. The no-reform payments trajectory runs 0.9% to 3.5% above the Treasury page 3 / page 4 implied no-reform totals. This is because the category-level reconciliation embeds the SDA Pricing Review trajectory and current SIL pricing growth observed in NDIA reporting, both of which trend slightly above Treasury’s aggregate no-reform projection.

Cohort reconciliation. The SIL/SDA cohort is sized from NDIA Q3 2025-26 data (SIL participants 36,808; SDA users 16,263; SDA eligible-not-using 9,370; estimated overlap-adjusted total approximately 41,265 at March 2026). The cohort trajectory grows to 51,654 by Jun 2031 in line with observed cohort growth rates.

A.4 Cohort incidence modelling Two cohort incidence scenarios are modelled, both calibrated to Treasury’s $38.1 billion 5-year savings envelope.

Scenario A — Cohort-neutral distribution. Each of the ten measures is mapped to specific support categories (or to all categories for measures with NDIS-wide application). For each measure-quarter,

Page 18

Submission 467

the SIL/SDA share of savings is set proportional to the cohort’s share of no-reform spending in the targeted categories.

Scenario B — SIL/SDA cohort protection. SIL/SDA participants are exempted from measures 1, 2, and

Cohort incidence is set to 0% for these measures. Other measures (M3, M4, M8, M9, M10) retain their proportional incidence as in Scenario A. Measure 6 (Thriving Kids) and Measure 7 (Functional Capacity Test) have cohort incidence of 0% by construction in both scenarios — SIL/SDA participants would not exit under either measure.

The two scenarios produce identical total savings at each FY anchor (Treasury’s $38.1 billion envelope is preserved), differing only in the allocation of the per-participant savings between cohorts.

A.5 Exit selection model The 241,000 stated exit cohort is allocated across projected cells using a vulnerability-weighted selection model.

For each cell (Disability × Age intersection in the non-SIL/SDA layer), an “eligibility-test vulnerability” weight is assigned, reflecting the likelihood that participants in that cell would fail the new Objective Functional Capacity Test. Weights range from 0.05 (cerebral palsy, down syndrome, spinal cord injury — disabilities with more severe and consistent functional impairment) to 0.65 (developmental delay — child-focused, Thriving Kids overlap). Within each disability, age-band weights are applied — 1.5 for 0-8 (Thriving Kids weight plus access tightening), 1.3 for 9-14, declining to 0.85 for adult age bands.

Per-cell exit allocations are computed as: pop × disability-vulnerability × age-vulnerability, normalised to sum to 241,000 across non-SIL/SDA cells. A 50% cap per cell (70% for high-vulnerability child cells aged 0-8 in developmental disabilities) prevents the model from over-allocating to small low-cost cells.

The model is calibrated to Treasury’s implied cumulative cost per exited participant of approximately $31,000 at the Jun 2031 anchor, achieved at $31,224 in the final model (within 1% of target).

The 105,000 inflow gap is allocated using the same vulnerability framework with heavier weight on 0 8 (weight 2.0) and 9-14 (weight 1.5) age bands to reflect the concentration of Thriving Kids effects and reduced new-entrant flows in younger participants.

A.6 Sensitivities and limitations Payment baseline. The no-reform payments run 0.9% to 3.5% above Treasury’s implied no-reform totals across the projection. To preserve reconciliation, all savings figures in the cohort scenarios are calibrated to Treasury’s published savings envelope. This means the percentage reductions reported (e.g. 15% for SIL/SDA under Scenario A) are reductions against the projected no-reform anchor, not against Treasury’s exact post-reform anchor. Reductions against Treasury’s no-reform would be marginally smaller (approximately 0.5-1 percentage point lower).

Cohort incidence mapping is assumption driven. Treasury has not disclosed which support categories each measure targets in detail or whether the operational application is uniform across cohorts. The support category mappings in Scenario A are reasoned interpretations of the measure titles and descriptions; they are not authoritative. Sensitivity testing (varying the support category mappings

Page 19

Submission 467

within plausible ranges) produces cohort-incidence variations of ±3 percentage points around the headline figures.

Exit selection is illustrative. The exit selection model produces a defensible projection of which cells contribute to the 241,000 exits, calibrated to match Treasury’s cumulative cost-per-exit anchor. The actual operational selection of which individual participants exit under the Objective Functional Capacity Test will be determined by the operational design of the test, which is not yet disclosed. The model’s output should be read as the most likely cohort and demographic profile of exits consistent with the published data, not as a prediction of which individuals will exit.

The 105,000 inflow gap is undocumented in the Treasury material. The analysis identifies this gap from the arithmetic of the published participant projection and offers a plausible mechanism (primarily reduced new-entrant flows attributable to measures 1 and 5). Without Treasury disclosure of the gap’s source, the model’s allocation of the 105,000 across cells is an inference.

Page 20

Submission 467